Initialise

Load Data

# Find files
samples.files <- grep("Results.*.csv", list.files(path = "."), value=TRUE)
rands.files <- grep("Rands.*.csv", list.files(path = "."), value=TRUE)

if (length(samples.files) != length(rands.files)) {
  stop("For some reason there are a different number of results files and random number files.
       This is weird so I'm not continuing.")
}

if (length(samples.files) != 1) {
  message("Found multiple samples files. Using the most recent one.")
  samples.files <- samples.files[-1]
  rand.files  <- rand.files[-1]
}

## Found multiple samples files. Using the most recent one.

## Error in eval(expr, envir, enclos): object 'rand.files' not found

# Read the file in
samples <- read_csv(samples.files[1], col_names=FALSE)

## Parsed with column specification:
## cols(
##   .default = col_integer(),
##   X101 = col_character()
## )

## See spec(...) for full column specifications.

rands <- read_csv(rands.files[1], col_names=FALSE)

## Parsed with column specification:
## cols(
##   .default = col_double(),
##   X21 = col_character()
## )
## See spec(...) for full column specifications.

samples[,ncol(samples)] <- NULL # Last column is NA because I lazily write too many commas
rands[,ncol(rands)] <- NULL # Last column is NA because I lazily write too many commas


# Truth data are the first entry
truth.results <- samples[1,]
samples <- samples[-1,]

truth.rands <- rands[1,]
rands <- rands[-1,]

Plot results (look at the people per iteration)

Spaghetti plot

# Define a fucntion to plot the samples
spag_plot <- function(samples, truth) {

  x_axis <- 1:ncol(samples)
  sample_col <- alpha("#ADD8E6", 0.9)
  truth_col <- "red"

  #create plot
  plot(1, xlim=c(0, max(x_axis)), ylim=c(min(rbind(truth,samples)), max(rbind(truth,samples))), type='l',
       xlab="Number of iterations", ylab = "Number of agents" )
  map(1:nrow(samples), function(x) lines(x_axis, samples[x,], type='l', col=sample_col))
  lines(x_axis, truth, type='l', lwd=2.5, col = truth_col)

  legend("bottomright", c("Samples", "Truth"), lty=c(1,1), lwd=c(2.5,2.5),col=c(sample_col, truth_col))
}


#map2(samples, obIntervals, function(x, obInterval) spag_plot(x, truth, obInterval))
spag_plot(samples, truth.results)

Density plot

Create the plots

# Returns a single ggplot2 density plot
density_plot <- function(s, truth) {

  # Y values for this sample (num people per iteration)
  y <- t(s[1,])[,1]
  # X values
  x <- 1:length(y)
  # A long two-column matrix will have number of people at every iteration across all samples
  scatter <- cbind(x, y)
  # Now add the rest of the samples
  for (i in 2:nrow(s)) {
    scatter <- rbind(scatter, cbind(1:length(y), t(s[i,])[,1]) )
  }
  scatter <- as.data.frame(scatter)

  the.plot <- ggplot(data=scatter, mapping=aes(x,y)) +
    geom_hex(bins=35, show.legend = FALSE) +
    geom_line(data = data.frame("x"=1:length(truth), "y"=t(truth)), mapping=aes(x,y), colour="red") + # Truth
    #geom_smooth(method="loess", se=TRUE, level=0.95, color="orange") +
    #geom_point(size=0.1, color="white") + # individual points
    ylab("Number of agents in the system")+
    xlab("Iteration")
    #ggtitle(paste("Observation Interval =", obs, sep =' '))

  return(the.plot)
}

#plots <- map2(samples, obIntervals, function(x, obInterval) density_plot(x, truth, obInterval))
density_plot(samples, truth.results)

Plot random numbers

XXXX Look at how the random numbers changed

(For now just make a table)

'True' random numbers

print(xtable(
  truth.rands
  ), type = "html")

Random number samples

print(xtable(
  rands
  ), type = "html")